Seismic Evidence for Large-Scale Compositional Heterogeneity of Oceanic Core Complexes

Home , Oceanic core complex

Article Geochemistry 3 Volume 9, Number 8 Geophysics 6 August 2008 Q08002, doi:10.1029/2008GC002009 GeosystemsG G ISSN: 1525-2027 AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

Click Here for Full Article Seismic evidence for large-scale compositional heterogeneity of oceanic core complexes

J. Pablo Canales and Brian E. Tucholke Department of Geology and Geophysics, Woods Hole Oceanographic Institution, MS 24, 360 Woods Hole Road, Woods Hole, Massachusetts 02543, USA ([email protected]) Min Xu Massachusetts Institute of Technology-Woods Hole Oceanographic Institution Joint Program, Woods Hole, Massachusetts 02543, USA John A. Collins and David L. DuBois Department of Geology and Geophysics, Woods Hole Oceanographic Institution, MS 24, 360 Woods Hole Road, Woods Hole, Massachusetts 02543, USA

[1] Long-lived detachment faults at mid-ocean ridges exhume deep-seated rocks to form oceanic core complexes (OCCs). Using large-offset (6 km) multichannel seismic data, we have derived two-dimensional seismic tomography models for three of the best developed OCCs on the Mid-Atlantic Ridge. Our results show that large lateral variations in P wave velocity occur within the upper 0.5–1.7 km of the lithosphere. We observe good correlations between velocity structure and lithology as documented by in situ geological samples and seafloor morphology, and we use these correlations to show that gabbros are heterogeneously distributed as large (tens to >100 km2) bodies within serpentinized peridotites. Neither the gabbros nor the serpentinites show any systematic distribution with respect to along-isochron position within the enclosing spreading segment, indicating that melt extraction from the mantle is not necessarily focused at segment centers, as has been commonly inferred. In the spreading direction, gabbros are consistently present toward the terminations of the detachment faults. This suggests enhanced magmatism during the late stage of OCC formation due either to natural variability in the magmatic cycle or to decompression melting during footwall exhumation. Heat introduced into the rift valley by flow and crystallization of this melt could weaken the axial lithosphere and result in formation of new faults, and it therefore may explain eventual abandonment of detachments that form OCCs. Detailed seismic studies of the kind described here, when constrained by seafloor morphology and geological samples, can distinguish between major lithological units such as volcanics, gabbros, and serpentinized peridotites at lateral scales of a few kilometers. Thus such studies have tremendous potential to elucidate the internal structure of the shallow lithosphere and to help us understand the tectonic and magmatic processes by which they were emplaced.

Components: 10,396 words, 13 figures. Keywords: oceanic core complex; detachment fault; Mid-Atlantic Ridge; seismic structure; gabbro; serpentinized peridotite. Index Terms: 3025 Marine Geology and Geophysics: Marine seismics (0935, 7294); 3035 Marine Geology and Geophysics: Midocean ridge processes; 3036 Marine Geology and Geophysics: Ocean drilling. Received 28 February 2008; Revised 10 June 2008; Accepted 27 June 2008; Published 6 August 2008.

Canales, J. P., B. E. Tucholke, M. Xu, J. A. Collins, and D. L. DuBois (2008), Seismic evidence for large-scale compositional heterogeneity of oceanic core complexes, Geochem. Geophys. Geosyst., 9, Q08002, doi:10.1029/2008GC002009.

1. Introduction (MAR): Kane [Dick et al., 2008; Tucholke et al., 1998], Dante’s Domes [Tucholke et al., 2001], and [2] Oceanic core complexes (OCCs) are deep sec- Atlantis [Cann et al., 1997] (Figure 1). The to- tions of the oceanic lithosphere exhumed to the mography models reveal large lateral variations in seafloor by long-lived detachment faults [Cann et P wave velocity within the shallow lithosphere al., 1997; MacLeod et al., 2002; Tucholke et al., beneath all three OCCs, as well as distinctly 1998]. When active, these faults constitute the sole different velocity structure in the associated volca- extensional boundary between separating tectonic nic hanging walls. A strong correlation between the plates [deMartin et al., 2007; Smith et al., 2006; modeled velocity variations and lithology of in situ Tucholke et al., 1998], and in some instances they geological samples allows us to constrain the large- can accommodate extension for 1–2 Ma forming scale spatial distribution of gabbroic intrusions and relatively smooth-surfaced, dome-shaped mega- serpentinized peridotites within the OCCs, thus mullions due to footwall rollover [Tucholke et al., providing valuable new insights into the formation 1996, 1998]. These features commonly exhibit and evolution of these features. spreading-parallel corrugations or ‘‘mullions’’ that can have amplitudes up to several hundred meters, 2. Geological Settings and they have been identified at a wide variety of spreading centers [Cannat et al., 2006; Ohara et [5] The three OCCs studied show characteristic al., 2001; Okino et al., 2004; Searle et al., 2003; smooth surfaces and spreading-parallel corruga- Tucholke et al., 1998, 2008]. tions typical of megamullions formed by major oceanic detachment faults [Tucholke et al., 1998]. [3] The abundance of gabbros and peridotites Hereinafter we use Tucholke et al.’s [1998] recovered in most OCCs [Blackman et al., 2002, definitions of breakaway zone (where the faults 2006; Cann et al., 1997; Dick et al., 2000, 2008; initially nucleated) and termination of the faults. Escartı´n et al., 2003; MacLeod et al., 2002; Reston Breakaway zones are generally defined by iso- et al., 2002; Tucholke and Lin, 1994], together with chron-parallel ridges that mark the older limit of gravity modeling that suggests the presence of fault corrugations; however, early corrugations mantle-like densities near the seafloor [e.g., often are not well developed and this limit can be Nooner et al., 2003], indicates that OCCs may be difficult to determine. Terminations are more easily ideal tectonic windows where the geological record identified because they mark the young limit of a of mantle flow and melt generation and migration smooth, corrugated fault surface against rougher can be studied [Tucholke et al., 1998]. They also adjacent terrain of the hanging wall. have the potential for hosting serpentinite-based hydrothermal systems [Fru¨h-Green et al., 2003; Kelley et al., 2001]. However, the relative abun- 2.1. Kane Oceanic Core Complex dances and distribution of these lithologies are not [6] Kane OCC is located between 30 and 55 km well constrained. Drilling at OCCs to date has off-axis on the North American plate immediately recovered dominantly gabbroic rocks [Blackman south of the Kane Transform Fault (TF), between et al., 2006; Dick et al., 2000; Kelemen et al., 23°200Nand23°400N (Figure 1). It formed 2004], raising a question of the extent to which between 3.3 Ma and 2.1 Ma under conditions of OCCs actually provide access to the oceanic man- strongly asymmetrical spreading (17.9 mm a1 on tle. Thus the origin and evolution of OCCs, their the OCC side to the west, and 7.9 mm a1 on the significance as tectonic windows into the oceanic conjugate between Chrons 2 and 2A [Williams et lithosphere, and their potential as widespread hosts al., 2006]). The breakaway for this OCC is marked of serpentinite-based hydrothermal ecosystems re- by a linear ridge that is continuous for more than main uncertain. 35 km along isochron (with the possible exception of a small offset near latitude 23°300N, Figure 1). [4] In this paper we present seismic tomography The detachment fault constituting the surface of the images obtained across three of the best developed OCC exhibits several domes (Babel, Abel, Cain, and best studied OCCs on the Mid-Atlantic Ridge

2of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 1. Shaded-relief, three-dimensional perspectives of multibeam bathymetry over Atlantis, Kane, and Dante’s Domes OCCs. (Note the different view angle of the Dante’s Dome image). Solid lines locate seismic profiles (numbered with A, DD, and K). Dashed lines show locations of breakaways and terminations of detachment faults. SR and CD are the Southern Ridge and Central Dome, respectively, of Atlantis OCC; ND and SD are the northern and southern domes, respectively, of Dante’s Domes OCC; and AD, BD, CD, and ED are the Abel, Babel, Cain, and Eve domes, respectively, of Kane OCC. Yellow star on Atlantis OCC locates the Lost City hydrothermal field. The inset at left shows the location of the three OCCs on the flanks of the MAR in the northern central Atlantic.

Adam, and Eve, following the nomenclature of serpentinized peridotite at the center of the OCC, Dick et al. [2008]) and is cut by a large-offset, providing direct evidence of the footwall compo- high-angle, west-facing normal fault (East Fault, sition there [Dick et al., 2008]. labeled in Figure 1). Extensive geological sampling indicates that the central and western parts of the 2.2. Dante’s Domes Oceanic Core Complex Kane OCC are predominantly ultramafic [Dick et [7] Dante’s Domes OCC lies just east of the MAR al., 2008]. Both peridotites and gabbros are ex- rift valley on the African plate, between 26 320N posed along the northern edge of the OCC (south ° and 26 400N, away from any major segment wall of the Kane TF) [Auzende et al., 1994]. Slide ° discontinuity [Tucholke et al., 2001] (Figure 1). scars along East Fault expose massive outcrops of Total spreading rate in the area during formation of

3of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G the OCC was 22 mm a1 [DeMets et al., 1990]. towed at a nominal depth of 10 m. Hydrophone A set of ridges near the eastern limit of the OCC group (i.e., receiver) spacing was 12.5 m. The defines a broad breakaway zone, located in 2.0– sound source was a tuned, 10-element air gun array 2.8 Ma lithosphere. To the west, the detachment with a total capacity of 51 L (3100 in3) triggered by surface is terminated by an oblique set of valleys distance every 37.5 m and towed at a nominal and ridges created by a rift that propagated south- depth of 8 m. Data were recorded in 10-s long ward between 1.3 and 0.7 Ma. The OCC exhibits records sampled every 4 ms. Positions of sources two domes that are adjacent along strike. Submers- and receivers were derived from shipboard and tail- ible dives show that dominantly allochthonous buoy GPS receivers and compass-enhanced basalt debris and minor serpentinite is scattered DigiCourse birds placed along the streamer. The across the corrugated detachment surface. There long-offset hydrophone streamer and the shallow are no direct constraints on composition of the depths of the OCCs allowed recording of subseafloor footwall, but gravity modeling indicates densities refractions that we use to derive the two-dimensional consistent with gabbros and/or partially serpenti- seismic velocity structure beneath the exposed nized peridotites [Tucholke et al., 2001]. detachment-faults surfaces using traveltime tomography methods. 2.3. Atlantis Oceanic Core Complex [10] Six profiles were acquired at the Kane OCC, [8] Atlantis OCC lies just west of the MAR rift three profiles at Dante’s Domes OCC, and five valley immediately north of Atlantis TF, between profiles across Atlantis OCC. In this paper we 30°050N and 30°170N (Figure 1), where the full present results from profiles that best sample the spreading rate is 24 mm a1 [Sempe´re´etal., main morphological features of the OCCs, i.e., two 1995]. The location of the breakaway is ambigu- profiles that cross each of the corrugated surfaces ous, but it is most likely located within a zone limited along dip and strike directions (Figure 1). Results by two north-south trending ridges (Figure 1) in from the complete set of the Kane profiles support 2.31.9 Ma lithosphere. The termination is marked the results presented here and provide a more by a contact between the smooth corrugated detach- detailed view of the Kane OCC [Xu et al., 2007]; ment surface and a basaltic hanging wall block in those results will be published elsewhere. The 0.7 Ma lithosphere that forms the western shoul- remaining profiles of the Atlantis OCC have yet deroftheriftvalley[Blackman et al., 2002; to be modeled and will be published elsewhere. Tucholke et al., 1998]. Atlantis OCC has two main morphological units [Blackman et al., 2002]: the [11] Shot gathers exhibit clear subseafloor refrac- Central Dome and the elevated Southern Ridge tions (Figure 2) and only minimum processing was along the top of the northern wall of Atlantis TF required for this study. Bad traces were edited, and (Figure 1). Spreading-parallel corrugations are best every five consecutive shots were gathered and the developed on the Central Dome. The southern part common-offset traces were stacked and averaged of the Central Dome was drilled at Hole U1309D to form 480-fold ‘‘super shot gathers’’ in order to during IODP Expeditions 304 and 305 (Figure 1) improve the signal-to-noise ratio (SNR). The super shot gathers were filtered in the frequency-wave and 1415 m of predominantly gabbroic rocks were recovered [Blackman et al., 2006; Ildefonse et al., number domain to remove low-frequency cable 2007]. In contrast, large slide scars along the south- noise. Traveltime picks were done in band-pass ern flank of the Southern Ridge, where the Lost City filtered super shot gathers with frequency bands of serpentinite-hosted hydrothermal system is located 3-5-15-30 Hz or 3-5-30-50 Hz, depending on the [Kelley et al., 2001], expose massive outcrops of data quality. Traveltime uncertainty was assigned serpentinized peridotite with lesser gabbro [Karson on the basis of the SNR in a 100-ms window et al., 2006]. before and after the pick [Zelt and Forsyth, 1994]. Although traveltimes were picked in all of the super shot gathers where subseafloor refrac- 3. Data and Methods tions were observed, traveltime picks were decimated in the shot and receiver domains before the 3.1. Data Acquisition and Processing inversion. Decimation of the traveltimes was done primarily because the complete data set has a higher [9] The multichannel seismic data used in this study were acquired in 2001 aboard R/V Maurice lateral sampling than can be resolved by traveltime Ewing (cruise EW0102) using a 6-km-long, tomography. Ray theory limits the lateral resolution 480-channel Syntron digital hydrophone streamer to the first Fresnel zone [e.g., Yilmaz, 1987]. In our

4of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 2. Representative shot gathers from profile K1 over the Kane OCC. Horizontal axes are distances from source to receivers along the streamer, and vertical axes are reduced two-way traveltime (reduction velocities of 4kms1 and 4.5 km s1, top and bottom, respectively). case, the first Fresnel zone at an average seafloor within each shot gather (i.e., effective receiver depth of 2.5 km is 250 m for a dominant frequency spacing of 125 m). The decimated set of tra of 30 Hz (deeper below the seafloor the Fresnel zone veltime picks is shown in Figure 3 as color maps increases as the dominant frequency in the data of traveltime versus shot number and source- decreases and the propagation length increases). receiver offset. Therefore we used traveltime picks from one out of every 5 shot gathers (i.e., effective source spacing [12] Feathering of the streamer was minimal and of 187.5 m), and from one out of every 10 receivers ranged between 6° (profile K7) and 3° (profiles

5of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 3. Color plots showing observed traveltime picks (two-way traveltime reduced at 3.7 km s1) displayed as a function of shot number and source-receiver offset for each of the seismic profiles. White areas indicate that no traveltimes were picked for a particular shot-receiver pair because of noisy data or because subseafloor refractions arrive later than the seafloor reflection at short offsets.

K1 and A4), resulting in receivers being off the used for raypath and traveltime calculations (see ideal two-dimensional profile by a maximum of next section). Perhaps the most important influence 650 m at the largest offsets. (Feathering was not of streamer feathering in our results is the three- possible to calculate for profile A10 owing to dimensional structure in the vicinity of the profiles corrupted readings by the streamer tail buoy that cannot be modeled with our approach. For GPS; for this profile we assumed that all of the example, the deepest parts of the tomography sources and receivers were perfectly aligned along models are most sensitive to feathering because the profile.) Projection of these out-of-plane re- they are constrained by the largest offsets; therefore ceiver locations into a 2-D profile does not intro- they probably represent an average structure within duce significant uncertainty in the observed a few hundred meters of the profile in the cross-line traveltimes because the differences between the direction. true and projected source-receiver offsets are less than 35 m, on the same order as the grid spacing

6of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

the same order as the first Fresnel zone at the seafloor (see previous sections) to avoid over- parameterizing the inverse problem. As a starting velocity model we used a laterally invariant model with constant vertical velocity gradient that best represents the one-dimensional structure of the Kane OCC [Xu et al., 2007] (3.7 km s1 at the seafloor and vertical gradient of 1.75 s1 below the seafloor, Figure 4). Regularization was achieved by minimizing an objective function dependent on the traveltime residuals (weighted by their uncertainty) and the second spatial partial derivatives of the model parameters [Zelt and Barton, 1998]. Readers should note that minimizing the second spatial derivatives of the model parameters results in models that have minimum roughness and therefore deviate as little as possible from the linear velocity gradient of the starting model.

[15] The trade-off between minimizing data residuals or obtaining a smooth solution is controlled by the damping parameter l [Zelt and Barton, 1998]. Figure 4. One-dimensional velocity structures at We ran inversions with different values of l, and at Atlantis OCC extracted from profile A4 beneath the Central Dome at the location closest to Hole U1309 each of the three study areas we selected the results that provided a weighted traveltime misfit function (blue) and beneath the Southern Ridge at 9.5 km 2 model distance (red). Solid black line shows sonic-log c 1.1 in the minimum number of iterations velocities measured in situ within the upper 200– (Figure 5). Our preferred solutions are shown in 800 m of IODP Hole U1309D [Blackman et al., 2006]. Figure 6, and they correspond to the second itera- Long-dash gray line shows the initial velocity model tion with l = 20 at the Atlantis and Dante’s Domes used in our tomography inversions. Short-dash blue line sites, and l = 10 at the Kane site (5 iterations were shows an alternate initial velocity model for profile A4 run in all cases). The importance of horizontal that was used to obtain the results shown in Figure 11. versus vertical smoothing constraints is controlled by the parameter sz [Zelt and Barton, 1998], with sz = 0 indicating no smoothing is imposed in the 3.2. Streamer Traveltime Tomography vertical direction, sz = 1 meaning roughness of model is equal both along vertical and horizontal 13 [ ] Traveltime tomography models were obtained directions, and sz > 1 resulting in models with using the software FAST [Zelt and Barton, 1998]. stronger smoothing constraints in the vertical The forward problem (ray tracing and traveltime direction. We found that our results were not calculation) was solved in a regular grid with 25 m significantly sensitive to values of sz < 0.075, while node spacing both vertically and horizontally. Be- larger values resulted in large c2 misfits, so we cause most of the subcrustal refractions observed in used sz = 0.075 in all of the inversions. our data are not true first arrivals (a common requirement in most traveltime tomography meth- [16] The strong lateral variations in velocity along ods) owing to the deep-water environment (first the profiles (Figure 6) are required by, and not arrivals in our data correspond to the direct wave simply consistent with, the observed data. For propagating horizontally from sources to receivers example, Figure 7 shows results for profile K1 through the water), the ray tracing algorithm was that include observed seismograms, traveltime modified in order to calculate raypaths and travel- picks, raypaths, and predicted traveltime curves times from subseafloor arrivals. for two representative shot gathers that sample very different velocity structures. The initial veloc- [14] The nonlinear inverse problem was solved ity model accurately predicts the observed travel- using a regularized, damped least squares solution. times for shot# 4516, but predicts traveltimes that The inverse model was parameterized as a regular are too large for shot# 4746, demonstrating the pres- grid with 200 m node spacing both vertically and ence of large lateral velocity variations. Figure 7 also horizontally. This grid spacing was chosen to be of

7of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

shows the complete ray coverage for profile K1. This example shows that ray coverage is dense and relatively homogeneous along the profiles.

[17] The large negative and positive traveltime residuals predicted by the initial velocity model are reduced to significantly lower residuals by the final models (Figure 8). Data are generally equally well fitted across all source-receiver offsets and shots, particularly for profiles K1, K7, and A10 (Figure 9), except at some locations where topography is rougher and/or seafloor deeper (e.g., near shot 5200 in profile K1, Figure 9). However, profiles A4, DD2, and DD4 show some depen- dence of traveltime residuals on source-receiver offset, with positive residuals appearing at near offsets and negative residuals at large offsets (Fig- ure 9). This indicates that although the preferred models provide a satisfactory statistical fit to the observations and their uncertainties, the models do not fully capture the complexity and small-scale subseafloor heterogeneities that are likely to be present beneath the OCCs. Thus, our results should be seen as one possible solution to the non-unique inverse problem, representing the minimum large- scale structure that is needed to fit the data within the imposed smoothing constraints. Other approaches such as those of Harding et al. [2007] that incorpo- rate into the inversion the traveltimes from the nearest offsets (not used in this study; see Figure 3), or high- resolution waveform tomography of streamer data [e.g., Shipp and Singh,2002],willbeneededto improve vertical resolution and better characterize the structure of OCCs, particularly within their shallow most (<200 m) parts.

3.3. Resolution Tests

[18] We conducted several tests to assess the lateral resolution of the final velocity models. We created synthetic velocity models by adding alternating positive and negative vertical anomalies to the initial model shown in Figure 4. The amplitudes of the anomalies varied laterally between ±0.5 km s1 following a sinusoidal function. Tests were done for anomalies with different widths (i.e., half wavelength of the sinusoidal perturbation) ranging Figure 5. Variation in misfit function between ob- from 1.5 km to 5.0 km. Synthetic traveltimes were served and predicted traveltimes as a function of calculated for each of the synthetic models using damping parameter l, for several inversion iterations. the same experimental geometry as in our seismic Gray squares indicate the preferred solutions shown in profiles. We then inverted the calculated travel- Figure 6. Black line corresponds to c2 =1.1,the times using the same parameters as in our tomog- adopted threshold value for the misfit function. raphy inversion. Only one iteration was required to achieve the desired misfit of c2 1.1. Results from profiles K1 and K7, which are representative

8of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 6. (a) Two-dimensional tomography models of P wave velocity beneath Kane, Dante’s Domes, and Atlantis OCCs (profile locations in Figure 1). Contours are every 0.5 km s1. Left panels are dip profiles, and right panels are strike profiles. Short-dashed lines on dip profiles indicate locations of breakaways and terminations. Long-dashed lines show crossings between dip and strike profiles. Green dots and blue diamonds (serpentinite and gabbro, respectively) show dominant in situ lithologies sampled by submersible or remotely operated vehicle [Auzende et al., 1994; Dick et al., 2008; Karson et al., 2006] or drilled [Blackman et al., 2006] (IODP Hole U1309D, vertical gray line projected onto profile A4). (b) Same as Figure 6a with models shown as perturbations with respect to the initial one-dimensional velocity model (Figure 4), with contours every 0.2 km s1.

9of22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 7. (a) Shot gathers from Figure 2 with observed traveltimes and their assigned uncertainty (circles with red error bars). Traveltime curves are shown as predicted by the initial model (green, from Figure 4) and the final model in Figure 6 (blue). (b) Final velocity model for profile K1, shown as perturbations with respect to the initial model, and contoured every 0.2 km s1. Raypaths (gray lines) for the two source-receiver pairs in Figure 7a, above, are shown. (c) Final velocity model for profile K1 with complete ray coverage and contours every 0.5 km s1.

10 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 8. Traveltime residual histograms. Results of both initial and final models are shown for each seismic profile. N indicates number of observations, RMS is root-mean squared, c2 is weighted misfit function, and mean is the mean of the distributions. for all the other profiles, are shown in Figure 10. this case the tests indicated that anomalies smaller The tests indicate that along sections of profile K1 than 2.5 km were not resolved in the final model. where seafloor is shallowest, features as small as 1.5 km wide in the final velocity model (Figure 6) 4. Results are meaningful and well resolved at subseafloor depths less than 0.5 km (Figure 10a). Along other 4.1. Seismic Structure parts of the profile where seafloor is deeper, only anomalies that are larger than 2.0–2.5 km wide [19] Our tomography models reveal strong lateral above 0.5 km subseafloor depth should be trusted gradients in P wave velocity at scales of 1 km or and interpreted. At subseafloor depths greater than less within the upper 0.5–1.7 km of the litho- 0.5 km, only features at scales of 5 km or larger sphere across the three OCCs (Figure 6). Velocities along profile K1 are meaningful. Similar results are everywhere greater than 3 km s1 but rarely were found for profile K7 (Figure 10b), although in exceed 6 km s1. We group velocity structures into three classes that commonly are separated by the

11 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 9. Color plots showing residual traveltimes predicted by the final models displayed as a function of shot number and source-receiver offset for each of the seismic profiles. transitions noted above. There are some deviations away zone (10 to 5 km model distance), and in from the bounds of the tripartite classification profile A10 from the Atlantis OCC breakaway described below, but the grouping captures the zone to the western flank of the Central Dome essentials of the observed variability. (13 to 6 km model distance).

[20] The first class includes areas with the lowest [21] The second class includes areas with interme- velocities near the seafloor (<3.4 km s1)and diate shallow velocities (3.4–4.2 km s1) and gen- also the lowest shallow vertical velocity gradients erally intermediate velocity gradients (1–3 s1). (<1s1); these are present in the three dip trans- These are observed beneath the Eve, Abel, and the ects (Figure 6a). These features appear in profile western half of Cain Dome at Kane OCC (profiles K1 east of the Kane OCC termination (volcanic K7 at 15 to 5 km, and K1 at 20 to 7 km), hanging wall terrane, 3 to 10 km model distance), beneath the northern dome of Dante’s Domes OCC in profile DD4 within the Dante’s Domes break- (profile DD2 at 5 to 7 km), and beneath the

12 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 10. Resolution tests for profiles (a) K1 and (b) K7 of the Kane OCC. Red-blue anomalies show recovered synthetic anomalies. Vertical long-dashed lines locate the alternating positive and negative true anomalies (widths indicated within each panel). Short-dashed lines correspond to a subseafloor depth of 0.5 km, except for the top panels, where it corresponds to 1 km subseafloor depth.

Southern Ridge and north of the Central Dome at anomalies, and the low-velocity areas by large Atlantis OCC (profile A4 at 13 to 7 km and 3 to negative anomalies. 9 km, respectively). 4.2. Comparison With Previous Gravity [22] The third class consists of areas with the highest shallow velocities (>4.2 km s1) and large Studies 1 shallow velocity gradients (>3 s ). These charac- [24] Gravity studies have been conducted for a teristics are observed beneath Babel Dome and the number of OCCs. Most of them have elevated eastern half of Cain Dome at Kane OCC (profiles residual mantle Bouguer anomalies, indicative of K7 at 5 to 15 km, and K1 at 6 to 0 km), beneath thin magmatic crust [e.g., Tucholke et al., 1998]. the southern dome at Dante’s Domes (profiles DD4 Density models derived from both surface and on- at 4 to13 km, and DD2 at 11 to 6 km), and bottom gravity measurements indicate that the beneath the eastern part of the Central Dome at cores of OCCs are characterized by relatively high Atlantis OCC (profiles A10 at 4 to 2 km, and A4 densities consistent with intrusive gabbros and/or at 7 to 2 km). partially serpentinized peridotite [Blackman et al., 1998, 2008; Matsumoto et al., 2001; Nooner et al., [23] These domains and the transitions between 2003; Tucholke et al., 2001]. In this section we them are best seen when velocities are displayed compare our seismic results with previously pub- relative to a laterally invariant, constant vertical- lished gravity-derived density models from the velocity-gradient model that was used as starting Atlantis and Dante’s Domes OCCs. We base our model for the tomographic inversions (Figure 4). comparison on the well established positive corre- As shown in Figure 6b, the high-velocity areas are lation between density and P wave velocity [e.g., marked by large positive anomalies, the intermediate- Carlson and Herrick, 1990] and discuss the simi- velocity areas by both small negative and positive larities and differences between our velocity mod-

13 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

This result conflicts with our finding of a funda- mentally different seismic structure between the Southern Ridge and Central Dome (profile A4, Figure 6b). Like any tomography inversion, our results depend to some degree on the initial assumptions (i.e., the initial velocity model). Therefore we explored the possibility that our results for profile A4 are influenced by the initial assumptions; for example, might data that sample the Southern Ridge be consistent with a broader range of plausible models than data that sample the Central Dome? We ran an alternate inversion of profile A4, using as a starting velocity model a 1-D Figure 11. Alternate solution for profile A4 across structure that is closer to the structure found Atlantis OCC shown as perturbations with respect to the beneath the Central Dome in our preferred solution initial one-dimensional velocity model labeled as ‘‘A4 1 alternate initial’’ in Figure 4; contours are every 0.2 km s1. (i.e., a seafloor velocity of 4.1 km s and vertical velocity gradient of 2.375 s1, Figure 4). This alternate model (Figure 11) converges to a trav- els and the existing density models, but we do not eltime misfit of c2 = 1.2 in 4 iterations. The results attempt here to interpret the density and velocity show the same general pattern as the preferred models in terms of lithology (complete interpreta- solution, i.e., high velocities beneath the Central tion of the models is discussed in section 5). Dome that are bounded to the north and south by significantly lower velocities. Note that compared 4.2.1. Atlantis to Figure 6b, Figure 11 shows larger-amplitude [25] Gravity data from Atlantis OCC indicate the negative anomalies and smaller-amplitude positive presence of a high-density core (2,900 kg m3) anomalies because the alternate starting model had beneath the Central Dome, distributed in a roughly higher velocity than in the preferred solutions. This triangular shape in east-west cross-section and test demonstrates that the contrasting seismic char- bounded to the east and west by lower-density acter between the Southern Ridge and Central material at shallow levels [Blackman et al., 1998, Dome is required by the data and is not biased 2002, 2008; Nooner et al., 2003]. The high-velocity by the starting model chosen. body observed on profile A10 (Figure 6) matches quite well the inferred shape of the high-density [27] To reconcile our velocity model for profile A4 core [e.g., see Blackman et al., 2008, Figure 4; with the gravity data, we are left with the possibil- Blackman et al., 2002, Figure 16; Nooner et al., ity that the model proposed by Blackman et al. 2003, Figure 3]; the west-dipping low- to high- [2008] may be simplistic in its use of a single density boundary approximately follows the low- density to define the core of the Southern Ridge. If, to high-velocity boundary and the bottom of the for example, the low velocities observed beneath low-velocity zone on the western side of the Central the Southern Ridge overlie high velocities (i.e., Dome. Also, the disappearance of the high-velocity unaltered peridotitic mantle) deeper than 1500– body toward the east near 1–2 km model distance 2000 m below seafloor, lithosphere beneath the coincides with eastward dip of the high-density core Southern Ridge could still have integrated density toward the axial valley and a transition to lower similar to that of the Central Dome, despite having densities at shallow levels. The independent seismic different structure (a layered structure or a vertical and density models thus are consistent with one density gradient) that would be consistent with our another and mutually support the idea of a core of seismic results. higher velocity and density basement beneath the eastern part of the Central Dome of Atlantis OCC. 4.2.2. Dante’s Domes [28] Density modeling across Dante’s Domes OCC [26] The most recent gravity analysis of Atlantis has been done using both surface and on-bottom OCC [Blackman et al., 2008] suggests that the gravity measurements [Tucholke et al., 2001]. The gravity signature of Atlantis OCC might be modeled density structure across this OCC from explained by the same average density beneath breakaway to termination agrees very well with our the Southern Ridge and Central Dome, with no seismic results. The eastern limit of the high- structural (i.e., density) contrast between them. velocity body near the breakaway zone (4km 14 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G model distance in profile DD4, Figure 6) coincides are exposed along the southern margin of the ridge with a density boundary between lower-density [Blackman et al., 2002; Karson et al., 2006] material to the east and the higher-density core of (Figures 6 and 12). In this area velocities of the OCC [see Tucholke et al., 2001, Figure 9]. The 6kms1 are not observed until 1500 m below density model also includes a high-density (i.e., the seafloor (Figures 4 and 6); this indicates large unaltered mantle) body located at 4.5 km beneath degrees of serpentinization (>80%) [Miller and the central part of the OCC that shoals to just Christensen, 1997] from the seafloor to at least 1.5 km near the breakaway. Unfortunately our those depths, in agreement with samples from the data do not image deeply enough to model this south flank of the Southern Ridge [Boschi et al., boundary, but future seismic experiments with a 2006]. Similar velocity structure occurs between larger source-receiver aperture could easily test the Abel and Cain Domes in the Kane OCC, where deeper parts of the density model. East Fault exposes massive outcrops of serpentinized peridotite [Dick et al., 2008]. The consisten- 5. Correlation of Seismic Structure and cy of this correlation indicates that serpentinized peridotites are predominant beneath the smooth, Lithology corrugated fault surface at Abel Dome and the western side of Cain Dome (Figures 6 and 12). [29] Extensive geological sampling has been con- This lithology-velocity correlation also suggests ducted at Atlantis and Kane OCC, and this allows that the intermediate-velocity area beneath the us to correlate the main seismic characteristics that northern dome at Dante’s Domes OCC (profile we observe with predominant lithologies. Our DD2) may be predominantly serpentinized perido- interpretation is also guided by considering the tite. A similar correlation may apply at Atlantis seismic results within their tectonic context and OCC north of the Central Dome at (profile A4), visualizing them together with the three-dimen- where gravity modeling indicates low-density sional surrounding topography (Figure 12 and basement rocks [Blackman et al., 2002, 2008]. auxiliary material1 Animations S1 and S2). However, seafloor morphology shows that volca- [30] At Atlantis OCC, velocities on profiles A4 and nic rocks form at least the upper part of the A10 closest to Hole U1309D at the Central Dome basement section there [Blackman et al., 2002, agree quite closely with the compressional veloc- 2008]. ities measured in situ within the hole [Blackman et [32] Areas with the lowest velocities most likely al., 2006] (Figure 4). We therefore argue that consist of volcanics and sheeted dikes, at least in gabbros are the dominant, although not necessarily their upper parts. These velocity characteristics exclusive, lithology forming the high-velocity appear in the hanging walls of the detachment body beneath the Central Dome. There is strikingly faults, as best seen on profile K1 where hummocky similar velocity structure in the other two OCCs seafloor morphology clearly shows the presence of near their terminations (Figures 6 and 12), and we dominant volcanics (Figure 12; see also auxiliary interpret all these high-velocity zones as represent- material Animations S1 and S2). They also occur ing predominantly gabbros. At Kane OCC, this is near the fault breakaways (Figure 6) where, for consistent with samples obtained from Babel Dome example at Dante’s Domes OCC, in situ pillow [Dick et al., 2008] and the southern wall of the basalts have been observed in submersible dives Kane TF [Auzende et al., 1994] (Figures 6 and 12), [Tucholke et al., 2001]. where gabbro appears to be the primary lithology. Our interpretation is also consistent with seismic [33] Our lithologic interpretation of the velocity and drilling results from ODP Hole 735B in the models indicates that there is large variability in Atlantis Bank OCC on the Southwest Indian the distribution of gabbros and serpentinized peri- Ridge; average seismic velocities of 6 km s1 dotites within the three OCCs studied here. At [Muller et al., 1997] are found there in the upper Atlantis OCC the Central Dome is mostly gabbroic crust where drilling recovered 1500 m of gab- from the summit of the corrugated surface to near broic rocks [Dick et al., 2000]. the termination. This gabbroic core is bounded to the south (and perhaps to the north) by predomi- [31] Intermediate velocities beneath the Southern nantly ultramafic lithosphere (Figures 6b and 12). Ridge at Atlantis OCC on the MAR correlate with Our seismic results clearly confirm a distinct lith- massive outcrops of serpentinized peridotite that ological contrast between the mafic Central Dome 1Auxiliary materials are available in the HTML. doi:10.1029/ and the ultramafic Southern Ridge, as previously 2008GC002009.

15 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 12. Three-dimensional renditions of seafloor topography (shaded gray) with intersecting vertical sections of the velocity profiles (red-blue images showing velocity perturbations as in Figure 6b, contoured every 0.2 km s1). Left panels show dip sections, and right panels show strike sections. Blue and red areas indicate velocities faster and slower than the reference model, respectively. Dashed lines on the seafloor topography indicate the limits of dominantly smooth, corrugated detachment surfaces; yellow segments show detachment terminations. Solid black lines draped on the seafloor locate the orthogonal seismic profiles at each OCC. Colored dots and diamonds are in situ lithologies as in Figure 6. Vertical yellow bar in profile A4 locates IODP Hole U1309D. Primary morphological features are labeled. Horizontal axes are latitude and longitude, and vertical axes show depth below sea level in meters. suggested on the basis of geological samples dominantly gabbroic composition appears to ex- [Karson et al., 2006]. tend nearly continuously from near the breakaway to the termination (Figures 6 and 12). The northern [34] At the Kane OCC, the velocity structure of dome, however, has velocity structure more like Cain Dome is very similar to the Atlantis Central that of the Southern Ridge at Atlantis OCC and Dome, with a gabbroic core forming the eastern Abel Dome at Kane OCC (Figure 12), suggesting a part of the dome, from the summit to near the predominance of serpentinites. termination (Figures 6 and 12). Farther north, the velocity structure of Babel Dome also indicates a gabbroic composition, but to the south Eve Dome 6. Discussion has an intermediate structure that may correspond to a mixture of mafic and ultramafic rocks (Figures 6 6.1. Distinguishing Between Gabbros and and 12). In contrast, the western side of Cain Dome Serpentinized Peridotites on the Basis of and the older Abel Dome are interpreted to be Their P Wave Velocity predominantly ultramafic. [36] Laboratory measurements of P wave velocity [35] At Dante’s Domes OCC the southern dome (VP) in hand-samples of oceanic gabbros and shows similarities to the Central Dome of Atlantis serpentinized peridotites at appropriate confining OCC and Cain Dome at Kane OCC, although a pressures indicate that VP ranges between 6.7 and

16 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

7.3 km s1 for gabbros, and between 4.8 and 8.0 the Atlantis OCC [Canales et al., 2004]. These km s1 for fully serpentinized and fresh peridotites, authors reported the presence of a shallow reflec- respectively [e.g., Miller and Christensen, 1997]. P tion, termed D, beneath parts of the core complex. wave velocity fields for both lithologies overlap Similar shallow reflections recently have been when the degree of serpentinization of the perido- reported at other oceanic core complexes [Ohara tites is 20–40%. For this reason there is inherent et al., 2007b]. The D reflection was best imaged ambiguity when interpreting as gabbros or serpen- along profile A10 (profile Meg-10 of Canales et al. tinized peridotites seismic velocities measured by [2004]), and was interpreted as a possible older remote methods (e.g., seismic refraction methods) detachment fault. IODP Hole U1309D cored in the 6.7–7.3 km s1 range, in the absence of through this reflection and recovered gabbros with other independent constraints such as direct sam- no evidence of a major fault interface; thus the pling, gravity data, or VP/VS ratio. In contrast, results do not support the interpretation of a sec- highly serpentinized peridotites (>40%) should be ondary detachment. Figure 13 compares our to- distinguishable from gabbroic rocks on the basis of mography results along profile A10 with the their VP, but they may become harder to differen- corresponding reflection image of Canales et al. tiate from basaltic rocks such as extrusive lavas and [2004], and it shows that the D reflection approx- dikes owing to their lower velocity. imately follows the 5.5 km s1 isovelocity con- tour, roughly marking the transition from low to [37] A key result from our study is that we can high velocities from west to east across the Central make such a distinction because (1) we base our Dome of Atlantis OCC. We suggest that the D interpretation on correlation to in situ samples and reflection at this location correlates with a rapid characteristic seafloor morphology, which allows increase in seismic velocity associated with the top us to distinguish volcanic settings from serpenti- of the gabbroic core forming the eastern half of the nized peridotite exposures, and (2) the sampled Central Dome. In this and adjacent areas where the serpentinized peridotites show high degrees of gabbroic core is probably close to, or exposed at, serpentinization (>70% [Boschi et al.,2006]), the seafloor (e.g., at the location of IODP Hole making them distinguishable from gabbros. The U1309D), the D reflection may correspond to high-velocity regions we interpret as predominant- velocity change related to a shallow fracturing or ly gabbros could also be interpreted as predomi- alteration front in the gabbro. nantly peridotites with moderate degree of serpentinization (20–40%); however, the constraints provided by the IODP Hole 1309D results 6.3. Lithospheric Structure During Early make this possibility unlikely. Stages of the OCC Formation [40] Differences among the three OCCs are em- [38] It is important to note that seismic velocities phasized in the velocity structure of the lithosphere derived from remote methods such as those used in exhumed shortly after fault breakaways (Figures 6 this study tend to be lower than velocities mea- and 12). The Kane detachment fault initially cut sured in laboratory samples in part because data through a shallow section of probable gabbro from the seismic experiments are sensitive to (Figure 6b) and shortly thereafter was exhuming fractures occurring at scales larger than the labo- mantle for about 0.5 Ma. This indicates that early ratory samples. This could explain why we mea- development of the Kane OCC was in a magma- sure velocities lower than 6 km s1 at the location limited environment. In contrast, the detachment at of IODP Hole 1309D (Figure 4), significantly the southern dome of Dante’s Domes OCC quickly lower than the 6.7–7.3 km s1 expected for intact exhumed a gabbro section that continued to fault gabbro samples [Miller and Christensen, 1997], or termination, while at Atlantis OCC the detachment velocities lower than 4.8 km s1 (the expected appears to have transected a thick, low-velocity, value for fully serpentinized peridotite) in the volcanic (and possibly dike?) section before upper 1 km of the Southern Ridge of Atlantis gabbros were finally exhumed. OCC (Figure 4).

6.2. Comparison of Velocity Model With 6.4. Large Gabbroic Intrusions and OCCs Seismic Reflectivity Beneath the Atlantis [41] The major common characteristic of each of OCC the three OCCs is the high-velocity body between the OCC summit and the termination, which we [39] Some of the data used in this study were used interpret as a gabbroic core (Figures 6 and 12). previously to image subseafloor reflections across

17 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Figure 13

18 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

Consistent with this apparent increased melt supply 90-km-long part of the OCC has lower seismic with time, all three OCCs show reduced residual velocities (6 km s1 are not reached until 2km mantle Bouguer gravity anomalies toward their below seafloor). This pattern of relatively high terminations [Nooner et al., 2003; Tucholke et al., seismic velocities near the termination is very sim- 1998, 2001]. It has been recently hypothesized that ilar to results along our dip profiles (particularly large gabbroic intrusions may be required to local- along profile K1), but it occurs at much larger ize strain and develop oceanic detachment faults scale. If the high-velocity body at Godzilla Mullion [Ildefonse et al., 2007] owing to the rheological was a gabbro core on which the detachment fault differences between gabbro and serpentinized pe- nucleated, it would require that the fault broke to ridotite [Escartı´n et al., 2001], and that the detach- an unrealistic depth of >60 km (following the same ments nucleate along the irregular brittle-plastic reasoning as above), which would argue against the transition around the gabbroic intrusions to form model of Ildefonse et al. [2007]. However, it must observed corrugations [Tucholke et al., 2008]. Our be noted that the Godzilla Mullion seismic dip line findings appear to support these ideas, although is only a narrow sample of velocity structure within there are some possible inconsistencies. Along the the 60-km wide OCC, and there could be high- Atlantis and Dante’s Domes dip profiles, the fault velocity gabbro bodies in lateral, along-strike surface from breakaway to the first downdip gab- directions near the breakaway that presently are bro body ranges in distance from 3–12 km; this undetected. would mean that the fault initially broke to depths [43] The strike lines at Dante’s Domes OCC and shallower than 11 km (assuming fault dip between 45° and 70°), which is not unreasonable Atlantic OCC (Figure 6b) show low velocities [deMartin et al., 2007]. A similar fault may have beneath the northern dome at Dante’s Domes and nucleated on the small, shallow body of probable beneath the Southern Ridge at Atlantis OCC. This gabbro near the breakaway of the Kane OCC observation also seems to argue against the model (profile K1, Figure 6b). However, if the Kane of Ildefonse et al. [2007], which predicts that the detachment instead nucleated above the gabbros Southern Ridge of Atlantis OCC is of the same in Cain dome, the fault would have extended nature/origin as the Central Dome, with a gabbroic core surrounded by a thin layer of serpentinized downdip 20 km and reached a depth of 14– 18 km, which seems unlikely. peridotite. However, it is possible that gabbro bodies are present closer to the breakaways, and [42] Similar distributions of apparent gabbro bod- they have not been detected in our seismic experi- ies near detachment terminations can be inferred at ments owing to the limited data coverage near Godzilla Mullion in the Parece Vela Basin in the breakaways. Thus, in order fully to understand Philippine Sea, which is the largest OCC found to the extent to which detachments may nucleate on date and occupies a 130-km-long, 60-km-wide deep-seated gabbro bodies, it will be necessary to corridor of corrugated seafloor between breakaway obtain strike-line velocity data in the critical zone and termination [Ohara et al.,2001].(Recent less than 10 km from detachment breakaways. seismic results from other OCCs in the Parece Vela Basin [Ohara et al., 2007b] are more difficult to [44] In contrast to the fault nucleation hypothesis compare to our results because those profiles are discussed above, an alternate interpretation for the oblique to the paleo-spreading direction and corru- gabbro cores near the terminations of the Kane, gations, and they do not always cross the main Dante’s Domes, and Atlantis OCCs is that they are domes of the OCCs.) An ocean bottom seismom- an independent, late-stage phenomenon. The en- eter seismic refraction profile in the dip direction hanced magmatism might occur in either of two over Godzilla Mullion [Ohara et al., 2007a] indi- ways. First, it could be part of a natural magmatic cycle controlled by the fertility or dynamics of melt cates that the 40-km-long section of the OCC near the termination is characterized by unusually flow in the upwelling mantle. Such cycles have high seismic velocities reaching 6 km s1 as been documented to occur in the flanks of the MAR at periods ranging from 0.4–0.8 Ma shallow as 500 m below seafloor, while the older, [Canales et al., 2000] to 2–3 Ma [Sempe´re´et

Figure 13. (a) Multichannel seismic reflection dip profile A10 across the top of the Central Dome of Atlantis OCC (profile Meg-10 in figure from Canales et al. [2004]) showing the shallow D reflection. The intersection of orthogonal profile A4 is indicated. (b) The same seismic section overlaid with the velocity tomography model of this study (labeled in km s1), lightly colored (color scale as in Figure 6a) and converted to two-way traveltime. (c) Seismic reflection profile as above but overlaid with velocity perturbation in km s1 relative to the starting model, as in Figure 6b. 19 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G al., 1995; Tucholke et al., 1997]. Alternatively, it in the OCCs. Such serpentinization can sustain might be the result of decompression melting in the hydrothermal circulation driven by exothermic exhuming footwall. Recent numerical models that reactions, as has been proposed for the Lost City explore the conditions that lead to the formation field on the Southern Ridge of Atlantis OCC and growth of long-lived oceanic detachment faults [Fru¨h-Green et al., 2003]. Thus, despite the com- predict that axial lithospheric thermal structure can mon occurrence of large gabbroic cores, OCCs, become highly asymmetrical owing to the advec- appear to have significant potential to host serpen- tion of hot material into the footwall of the fault tinite-based hydrothermal systems. Seismic studies [Tucholke et al., 2008]. Such asymmetrical thermal like those described here can be an important tool structure might eventually stimulate melting in the in the search for new non-magmatic hydrothermal exhuming footwall if the rising mantle is suffi- ecosystems within OCCs. ciently fertile, and it would thus result in emplacement of the gabbro bodies. Lateral variations in [49] We consistently observe high-velocity gabbros composition and fertility of the mantle could mod- toward the terminations of the OCCs studied. ulate melt production and thus account for the These might be emplaced either because of natural broader spatial heterogeneity of the gabbros. variability in the magmatic cycle or because of decompression melting within the rapidly exhum- [45] Irrespective of the cause of the increased melt ing footwall [Tucholke et al., 2008], but in either supply, a possible consequence is that this increase case the newly introduced heat may contribute to could lead to abandonment of the detachment fault formation of new faults and abandonment of the [Buck et al., 2005; Tucholke et al., 1998, 2008]. detachment. If decompression melting is signifi- Heat introduced by the flow and crystallization of cant, then large parts of OCCs, particularly those melt could weaken the lithosphere at the ridge axis close to the termination, may be affected by to the point that a new fault or set of faults would melting, flow, and emplacement of magma that form there and take up plate separation formerly overprints the pre-existing geological record of accommodated by the detachment. these processes. Although this kind of overprinting is to be expected in any ocean crust affected by the 7. Concluding Remarks flow of melt, the process may well be exaggerated in the case where rapid exhumation is caused by [46] In this study we have shown that seismic detachment faulting. reflection/refraction data acquired with a large- [50] If the gabbroic cores interpreted in our seismic offset hydrophone streamer is a powerful tool for profiles are an integral component of detachment determining the uppermost seismic structure of the faulting and OCC formation, then large oceanic oceanic lithosphere where it is exposed in relatively detachment faults are non-conservative because shallow OCCs. Such seismic studies, when inte- new material is added to the footwall during grated with good sample and morphological con- exhumation. For this reason, Dick et al. [2008] straints, can map with a high degree of confidence have proposed that such faults should be termed the lateral variations in the dominant composi ‘‘plutonic growth faults.’’ This hypothesis will tion of the oceanic lithosphere at scales of a need to be tested by thermal and mechanical few kilometers. numerical models such as those of Buck et al. [47] Our seismic tomography results indicate that [2005] and Tucholke et al. [2008]. 2 gabbros occur in large bodies (tens to >100 km ) [51] Finally, we note that our study is limited to the and have a heterogeneous distribution within shal- uppermost lithosphere, and it leaves open questions low lithosphere that is exhumed by long-lived about the deeper geometry of the gabbroic cores, oceanic detachment faults. This observation sug- along-strike variability near breakaways, and depth gests that large-scale melt flux from the mantle is extent of serpentinization. These questions must be irregularly distributed within spreading segments addressed with both deeper and more complete and is not necessarily focused at segment centers as geophysical imaging and by drilling. has been previously postulated [Whitehead et al., 1984]. A similar conclusion has been reached for Acknowledgments the Kane OCC on the basis of extensive rock sampling [Dick et al., 2008]. [52] This research was supported by grants from the U.S. NSF-IODP Program. We thank the captain, crew, and scien- [48] The tomography profiles also suggest widespread distribution of serpentinized peridotite with-

20 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G tific party of R/V Maurice Ewing Cruise 0102. This study and age, J. Geophys. Res., 95, 9153–9170, doi:10.1029/ benefited from discussions with H. Dick, J. Escartıń, and JB095iB06p09153. H. Schouten and from reviews by D. Blackman and an deMartin, B. J., R. Reves-Sohn, J. P. Canales, and S. E. anonymous reviewer. We thank H. Dick and M. A. Tivey Humphris (2007), Kinematics and geometry of active detachment faulting beneath the Trans-Atlantic Geotraverse for sharing data on Kane OCC geological samples with us (TAG) hydrothermal field on the Mid-Atlantic Ridge, Geol- during the course of our study. The software Focus by TM ogy, 35, 711–714, doi:10.1130/G23718A.1. Paradigm was used in the seismic data processing. DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein (1990), Current plate motions, Geophys. J. Int., 101, 425–478, References doi:10.1111/j.1365-246X.1990.tb06579.x. Dick, H. J. B., et al. (2000), A long in situ section of the lower Auzende, J.-M., M. Cannat, P. Gente, J.-P. Henriet, T. Juteau, ocean crust: Results of ODP Leg 176 drilling at the South- J. A. Karson, Y. Lagabrielle, C. Me´vel, and M. A. Tivey west Indian ridge, Earth Planet. Sci. Lett., 179, 31–51, (1994), Observation of sections of oceanic crust and mantle doi:10.1016/S0012-821X(00)00102-3. cropping out on the southern wall of Kane FZ (N. Atlantic), Dick, H. J. B., M. A. Tivey, and B. E. Tucholke (2008), Plu- Terra Nova, 6, 143–148, doi:10.1111/j.1365-3121. tonic foundation of a slow-spreading ridge segment: Oceanic 1994.tb00647.x. core complex at Kane Megamullion, 23°300N, 45°200W, Blackman, D. K., J. R. Cann, B. Janssen, and D. K. Smith Geochem. Geophys. Geosyst., 9, Q05014, doi:10.1029/ (1998), Origin of extensional core complexes: Evidence 2007GC001645. from the Mid-Atlantic Ridge at Atlantis Fracture Zone, Escartıń, J., G. Hirth, and B. Evans (2001), Strength of slightly J. Geophys. Res., 103(B9), 21,315–21,333, doi:10.1029/ serpentinized peridotites: Implications for the tectonics of 98JB01756. oceanic lithosphere, Geology, 29, 1023–1026, doi:10.1130/ Blackman, D. K., et al. (2002), Geology of the Atlantis Massif 0091-7613(2001)029<1023:SOSSPI>2.0.CO;2. (Mid-Atlantic Ridge, 30°N): Implications for the evolution Escartıń, J., C. Me´vel, C. J. MacLeod, and A. M. McCaig of an ultramafic oceanic core complex, Mar. Geophys. Res., (2003), Constraints on deformation conditions and the origin 23, 443–469, doi:10.1023/B:MARI.0000018232.14085.75. of oceanic detachments: The Mid-Atlantic Ridge core com- Blackman, D. K., B. Ildefonse, B. E. John, Y. Ohara, D. J. plex at 15°450N, Geochem. Geophys. Geosyst., 4(8), 1067, Miller, C. J. MacLeod, and Expedition 304/305 Scientists doi:10.1029/2002GC000472. (2006), Oceanic core complex formation, Atlantis Massif, Fru¨h-Green, G. L., D. S. Kelley, S. M. Bernasconi, J. A. in Proc. Integrated Ocean Drill. Program, 304/305, Karson, K. A. Ludwig, D. A. Butterfield, C. Boschi, and doi:10.2204/iodp.proc.304305.302006. G. Proskurowski (2003), 30,000 years of hydrothermal Blackman, D. K., G. D. Karner, and R. C. Searle (2008), activity at the Lost City vent field, Science, 301, 495– Three-dimensional structure of oceanic core complexes: 498, doi:10.1126/science.1085582. Effects on gravity signature and ridge flank morphology, Harding, A. J., G. M. Kent, D. K. Blackman, S. Singh, and J. P. Mid-Atlantic Ridge, 30°N, Geochem. Geophys. Geosyst., Canales (2007), A new method for MCS refraction data 9, Q06007, doi:10.1029/2008GC001951. analysis of the uppermost section at a Mid-Atlantic Ridge Boschi, C., G. L. Fru¨h-Green, A. Delacour, J. A. Karson, and core complex, Eos Trans. AGU, 88(52), Fall Meet. Suppl., D. S. Kelley (2006), Mass transfer and fluid flow during Abstract S12A-03. detachment faulting and development of an oceanic core Ildefonse, B., D. K. Blackman, B. E. John, Y. Ohara, D. J. complex, Atlantis Massif (MAR 30°N), Geochem. Geophys. Miller, C. J. MacLeod, and Integrated Ocean Drilling Pro- Geosyst., 7, Q01004, doi:10.1029/2005GC001074. gram Expeditions 304/305 Science Party (2007), Oceanic Buck, W. R., L. L. Lavier, and A. N. B. Poliakov (2005), core complexes and crustal accretion at slow-spreading Modes of faulting at mid-ocean ridges, Nature, 434, 719– ridges, Geology, 35, 623–626, doi:10.1130/G23531A.1. 723, doi:10.1038/nature03358. Karson, J. A., G. L. Fru¨h-Green, D. S. Kelley, E. A. Williams, Canales, J. P., J. A. Collins, J. Escartıń, and R. S. Detrick D. R. Yoerger, and M. Jakuba (2006), Detachment shear (2000), Seismic structure across the rift valley of the Mid- zone of the Atlantis Massif core complex, Mid-Atlantic Atlantic Ridge at 23°200 (MARK area): Implications for Ridge, 30°N, Geochem. Geophys. Geosyst., 7, Q06016, crustal accretion processes at slow spreading ridges, J. Geo- doi:10.1029/2005GC001109. phys. Res., 105(B12), 28,411–28,425, doi:10.1029/ Kelemen, P. B., E. Kikawa, D. J. Miller, and the Shipboard 2000JB900301. Scientific Party (2004), Proceedings of the Ocean Drilling Canales, J. P., B. E. Tucholke, and J. A. Collins (2004), Seis- Program, Initial Reports, vol. 209, Ocean Drill. Program, mic reflection imaging of an oceanic detachment fault: College Station, Tex., doi:10.2973/odp.proc.ir.2209.2004. Atlantis megamullion (Mid-Atlantic Ridge, 30°100N), Earth Kelley, D. S., et al. (2001), An off-axis hydrothermal vent field Planet. Sci. Lett., 222, 543–560, doi:10.1016/ near the Mid-Atlantic Ridge at 30°N, Nature, 412, 145–149, j.epsl.2004.02.023. doi:10.1038/35084000. Cann, J. R., D. K. Blackman, D. K. Smith, E. McAllister, MacLeod, C. J., et al. (2002), Direct geological evidence for B. Janssen, S. Mello, E. Avgerinos, A. R. Pascoe, and oceanic detachment faulting: The Mid-Atlantic Ridge, J. Escartıń (1997), Corrugated slip surfaces formed at ridge- 15°450N, Geology, 30, 879–882, doi:10.1130/0091- transform intersections on the Mid-Atlantic Ridge, Nature, 7613(2002)030<0879:DGEFOD>2.0.CO;2. 385, 329–332, doi:10.1038/385329a0. Matsumoto, T., A. Hosford, and H. J. B. Dick (2001), Gravity Cannat, M., V. Mendel, E. Ruellan, K. Okino, J. Escartıń, signal of the mantle exposure on the Atlantis Bank and the V. Combier, and M. Baala (2006), Modes of seafloor Atlantis-II fracture zone, Southwest Indian ridge (abstract), generation at a melt-poor ultraslow-spreading ridge, Geol- paper presented at EGS XXVI General Assembly, Eur. Geo- ogy, 34, 605–608, doi:10.1130/G22486.1. phys. Soc., Nice, France. Carlson, R. L., and C. N. Herrick (1990), Densities and por- Miller, D. J., and N. I. Christensen (1997), Seismic velocities osities in the oceanic crust and their variations with depth of lower crustal and upper mantle rocks from the slow

21 of 22 Geochemistry 3 canales et al.: heterogeneity of oceanic core complexes Geophysics 10.1029/2008GC002009 Geosystems G

spreading Mid-Atlantic Ridge, south of the Kane transform 13°N on the Mid-Atlantic Ridge, Nature, 442, 440–443, zone (MARK), Proc. Ocean Drill. Program Sci. Results, doi:10.1038/nature04950. 153, 437–454. Tucholke, B. E., and J. Lin (1994), A geological model for Muller, M. R., C. J. Robinson, T. A. Minshull, R. S. White, the structure of ridge segments in slow spreading ocean and M. J. Bickle (1997), Thin crust beneath ocean drilling crust, J. Geophys. Res., 99, 11,937–11,958, doi:10.1029/ program borehole 735B at the Southwest Indian Ridge?, 94JB00338. Earth Planet. Sci. Lett., 148, 93–107, doi:10.1016/S0012- Tucholke, B. E., J. Lin, and M. C. Kleinrock (1996), Mullions, 821X(97)00030-7. megamullions, and metamorphic core complexes on the Nooner, S. L., G. S. Sasagawa, D. K. Blackman, and M. A. Mid-Atlantic Ridge, Eos Trans. AGU, 77(46), Fall Meet. Zumberge (2003), Structure of oceanic core complexes: Suppl., F724. Constraints from seafloor gravity measurements made at Tucholke, B. E., J. Lin, M. C. Kleinrock, M. A. Tivey, T. B. the Atlantis Massif, Geophys. Res. Lett., 30(8), 1446, Reed, J. A. Goff, and G. E. Jaroslow (1997), Segmentation doi:10.1029/2003GL017126. and crustal structure of the western Mid-Atlantic Ridge Ohara, Y., T. Yoshida, and S. Kasuga (2001), Giant megamul- flank, 25°250-27°100 and 0-29 m.y., J. Geophys. Res., 102, lion in the Parece Vela Backarc basin, Mar. Geophys. Res., 10,203–10,223, doi:10.1029/96JB03896. 22, 47–61, doi:10.1023/A:1004818225642. Tucholke, B. E., J. Lin, and M. C. Kleinrock (1998), Mega- Ohara, Y., Y. Ohara, K. Okino, J. E. Snow, and Science Party mullions and mullion structure defining oceanic meta- (2007a), IODP site survey for drilling Godzilla Mullion: morphic core complexes on the Mid-Atlantic Ridge, Preliminary report of R/V Hakuho KH07-2 Leg 2 & 4 cruise, J. Geophys. Res., 103, 9857–9866, doi:10.1029/98JB00167. Eos Trans. AGU, 88(52), Fall Meet. Suppl., Abstract T53B- Tucholke, B. E., K. Fujioka, T. Ishihara, G. Hirth, and 1314. M. Kinoshita (2001), Submersible study of an oceanic Ohara, Y., K. Okino, and J. Kasahara (2007b), Seismic study megamullion in the central North Atlantic, J. Geophys. on oceanic core complexes in the Parece Vela back-arc basin, Res., 106, 16,145–16,161, doi:10.1029/2001JB000373. Island Arc, 16, 348–360, doi:10.1111/j.1440-1738. Tucholke, B. E., M. D. Behn, R. Buck, and J. Lin (2008), The 2007.00591.x. role of melt supply in oceanic detachment faulting and for- Okino, K., K. Matsuda, D. M. Christie, Y.Nogi, and K. Koizumi mation of megamullions, Geology, 36, 455–458, (2004), Development of oceanic detachment and asym- doi:10.1130/G24639A.1. metric spreading at the Australian-Antarctic Discordance, Whitehead, J. A., H. J. B. Dick, and H. Schouten (1984), A Geochem. Geophys. Geosyst., 5, Q12012, doi:10.1029/ mechanism for magmatic accretion under spreading centers, 2004GC000793. Nature, 312, 146–148, doi:10.1038/312146a0. Reston, T. J., W. Weinrebe, I. Grevemeyer, E. R. Flueh, N. C. Williams, C. M., M. A. Tivey, and M. D. Behn (2006), The Mitchell, L. Kirstein, C. Kopp, and H. Kopp (2002), A rifted magnetic structure of Kane megamullion: Results from mar- inside corner massif on the Mid-Atlantic Ridge at 5°S, Earth ine magnetic anomalies, paleomagnetic data and thermal Planet. Sci. Lett., 200, 255–269, doi:10.1016/S0012- modeling, Eos Trans. AGU, 87(52), Fall Meet. Suppl., 821X(02)00636-2. Abstract T42A-03. Searle, R. C., M. Cannat, K. Fujioka, C. Me´vel, H. Fujimoto, Xu, M., J. P. Canales, B. E. Tucholke, and D. L. Dubois A. Bralee, and L. Parson (2003), FUJI Dome: A large de- (2007), Shallow seismic structure of the Kane core complex, tachment fault near 64°E on the very slow-spreading south- Mid-Atlantic Ridge 23°300N, Eos Trans. AGU, 88(52), Fall west Indian Ridge, Geochem. Geophys. Geosyst., 4(8), 9105, Meet. Suppl., Abstract T53B-1304. doi:10.1029/2003GC000519. Yilmaz, O¨ . (1987), Seismic Data Processing, Invest. Geophys., Sempe´re´, J.-C., et al. (1995), The Mid-Atlantic Ridge between vol. 2, 526 pp., Soc. of Explor. Geophys., Tulsa, Okla. 29°N and 31°300N in the last 10 Ma, Earth Planet. Sci. Lett., Zelt, C. A., and P. J. Barton (1998), Three-dimensional seismic 130, 45–55, doi:10.1016/0012-821X(94)00259-2. refraction tomography: A comparison of two methods ap- Shipp, R. M., and S. C. Singh (2002), Two-dimensional full plied to data from the Faeroe Basin, J. Geophys. Res., 103, wavefield inversion of wide-aperture marine seismic strea- 7187–7210, doi:10.1029/97JB03536. mer data, Geophys. J. Int., 151, 325–344, doi:10.1046/ Zelt, C. A., and D. A. Forsyth (1994), Modeling wide-angle j.1365-246X.2002.01645.x. seismic data for crustal structure: Southeastern Grenville Pro- Smith, D. K., J. R. Cann, and J. Escartı´n (2006), Widespread vince, J. Geophys. Res., 99, 11,687–11,704, doi:10.1029/ active detachment faulting and core complex formation near 93JB02764.

22 of 22